lowrank_kernel-class: An S4 class to represent a low-rank kernel for a SNPset at...

In jmanitz/kangar00: Kernel Approaches for Nonlinear Genetic Association Regression

An S4 class to represent a low-rank kernel for a SNPset at specified knots

Description

An S4 class to represent a low-rank kernel for a SNPset at specified knots

Details

This kernel is used for predictions. If observations and knots are equal, better construct a full-rank kernel of class kernel.

Slots

type

character, kernel type: Use 'lin' for the linear kernel, 'sia' for the size-adjusted or 'net' for the network-based kernel.

kernel

kernel matrix of dimension equal to individuals

pathway

pathway object

Author(s)

Juliane Manitz

Examples

data(gwas)
data(hsa04020)
square <- calc_kernel(gwas, hsa04020, knots=gwas, type='lin', calculation='cpu')
dim(square@kernel)

gwas2 <- new('GWASdata', pheno=pheno[1:10,], geno=geno[1:10,], anno=anno, desc="study 2")
low_rank <- calc_kernel(gwas, hsa04020, knots = gwas2, type='net', calculation='cpu')
dim(low_rank@kernel)

jmanitz/kangar00 documentation built on Jan. 7, 2023, 11:26 a.m.